The purpose of this study is to examine whether measures of increased noradrenergic activity in haloperidol treated schizophrenic patients predicts relapse following drug withdrawal. Establishing predictors of rapid relapse following neuroleptic discontinuance will be important in targeted or intermittent neuroleptic treatment. One hundred carefully diagnosed, physically healthy, male schizophrenics will be recruited for the study. Subjects will be clinically stabilized on oral haloperidol prior to entering the study. After giving consent, each subject will receive a standard diet, and be evaluated with rating scales, as well as a series of biological measures related to noradrenergic function (three consecutive 24 hour urines for norepinephrine (NE), 3-methoxy-4-hydroxyphenolglycol (MHPG), vanillyl mandlic acid (VMA), normetanephrine (NM); a lumbar puncture for CSF NE and MHPG; and an intravenous clonidine challenge test to measure the plasma growth hormone (GH) and MHPG response). Subjects will then be withdrawn from haloperidol and followed as outpatients until relapse or for one year if they remain clinically stable. Patients will be assessed weekly to determine relapse and clinical stability. Data will be analyzed to test the hypothesis that noradrenergic variables measured during haloperidol maintenance treatment are predictors of subject's time to relapse. "Survival" analysis (Cox regression) will be used. This method of data analysis permits factors such as age, duration of illness, haloperidol blood levels, tardive dyskinesia, and the presence of brain atrophy to be considered.